1985 WL 669161 FOR EDUCATIONAL USE ONLY Page 1 1985 WL ... · FN1. With respect to the impact of Adams v. Texas, 448 U.S. 38 (1980), and Wainwright v. Witt, 105 S. Ct. 844 (1985),

For opinion see 106 S.Ct. 1758

Briefs and Other Related Documents

<<Material appearing in DIGEST section, including Topic and Key Number

classifications, Copyright 2006 West Publishing Company>>

Supreme Court of the United States.

A. L. LOCKHART, Director Arkansas Department of Correction, Petitioner,

v.

Ardia V. MCCREE, Respondent.

No. 84-1865.

October Term, 1985.

December 23, 1985.

On Writ of Certiorari to the United States Court of Appeals for the Eighth Circuit

Brief for Amicus Curiae American Psychological Association in Support of

Respondent

Donald N. Bersoff, (Counsel of Record), David W. Ogden, Ennis, Friedman, Bersoff & Ewing,

1200 17th Street, N.W., Suite 400, Washington, D.C. 20036, (202) 775-8100, Attorneys for

Amicus Curiae, American Psychological Association

*i TABLE OF CONTENTS

TABLE OF AUTHORITIES ... iii

INTEREST OF AMICUS CURIAE ... 1

INTRODUCTION AND SUMMARY OF ARGUMENT ... 2

ARGUMENT ... 4

I. THE SOCIAL SCIENCE DATA TENDERED BY RESPONDENT DEMONSTRATE THAT DEATH-QUALIFIED JURIES

ARE MORE PRO-PROSECUTION AND UNREPRESENTATIVE THAN TYPICAL CRIMINAL JURIES AND THAT DEATH

QUALIFICATION IMPAIRS JURY FUNCTIONING ... 4

A. The Social Science Research has Focused on the Relevant Categories of Prospective

Jurors, Including that Subset of Jurors Excludable Under Witherspoon ... 4

B. Social Science Research Conducted Over the Course of Three Decades Directly Addresses

the Constitutional Issues at Stake ... 5

1. The data show that death-qualified juries are conviction prone ... 6

2. The data show barring of "Witherspoon Excludables" creates unrepresentative juries,

thereby implicating defendant's right to a jury composed from a fair cross-section of the

community ... 16

1985 WL 669161 FOR EDUCATIONAL USE ONLY Page 1

1985 WL 669161 (U.S.)

© 2006 Thomson/West. No Claim to Orig. U.S. Govt. Works.

http://www.westlaw.com/Find/Default.wl?rs=dfa1.0&vr=2.0&FindType=Y&SerialNum=1986123246

a. Death qualification excludes a significantly large subset of the population ... 16

b. Those excluded through death qualification share common attitudes on issues related

to criminal justice ... 17

c. Death qualification results in under-representation on juries of blacks and women

... 19

*ii 3. The data suggest that death qualification interferes with the proper functioning

of the jury ... 20

II. CONTRARY TO THE STATES' CRITICISMS, WHICH ARE EITHER ERRONEOUS OR UNRELATED TO THE

PERTINENT RESEARCH, THE SOCIAL SCIENCE DATA TENDERED BY RESPONDENT SATISFY APPLICABLE

CRITERIA FOR EVALUATING THE SOUNDNESS OF SCIENTIFIC RESEARCH ... 21

CONCLUSION ... 30

*iii TABLE OF AUTHORITIES

CASES:

Adams v. Texas, 448 U.S. 38 (1980) ... 2, 5, 28, 29

Addington v. Texas, 441 U.S. 418 (1979) ... 21

Alexander v. Louisiana, 404 U.S. 625 (1972) ... 23

Apodaca v. Oregon, 406 U.S. 404 (1972) ... 20, 23

Ballew v. Georgia, 435 U.S. 223 (1978) ... 6, 9, 20, 23

Barefoot v. Estelle, 463 U.S. 880 (1983) ... 29

Carter v. Jury Commission, 396 U.S. 320 (1970) ... 23

Castenada v. Partida, 430 U.S. 482 (1977) ... 22, 23

Colgrove v. Battin, 413 U.S. 149 (1973) ... 23

Duren v. Missouri, 439 U.S. 357 (1979) ... 6

Grigsby v. Mabry, 569 F. Supp. 1273 (E.D. Ark. 1983), aff'd, 758 F.2d 226 (8th Cir. 1985)

... 4, 16, 17

Hazelwood Sch. Dist. v. United States, 433 U.S. 299 (1977) ... 22

Hovey v. Superior Court, 28 Cal. 3d 1, 616 P.2d 1301, 168 Cal. Rptr. 128 (1980) ... 4,

15, 19


1985 WL 669161 (U.S.)


http://www.westlaw.com/Find/Default.wl?rs=dfa1.0&vr=2.0&DB=780&FindType=Y&SerialNum=1980116796













In re Winship, 397 U.S. 358 (1970) ... 21

Johnson v. Louisiana, 406 U.S. 356 (1972) ... 20, 23

Peters v. Kiff, 407 U.S. 493 (1972) ... 20

Taylor v. Louisiana, 419 U.S. 522 (1975) ... 6, 20

Wainwright v. Witt, 105 S. Ct. 844 (1985) ... 2, 5, 14, 29

Williams v. Florida, 399 U.S. 78 (1970) ... 23

Witherspoon v. Illinois, 391 U.S. 510 (1968) ... passim

UNITED STATES CONSTITUTION:

Amendment VI ... 5, 16

Amendment XIV ... 5

MISCELLANEOUS:

Abelson, Three Models of Attitude-Behavior Consistency in 2 VARIABILITY AND CONSISTENCY

OF SOCIAL BEHAVIOR: THE ONTARIO SYMPOSIUM 131 (M. Zanna, C. Herman, E. Higgins eds. 1982)

... 24

ARKANSAS ARCHIVAL STUDY (1981) ... 16, 17, 27

D. BARNES. STATISTICS AS PROOF: FUNDAMENTALS OF QUANTITATIVE EVIDENCE (1983) ... 23

Berry, Death-Qualification and the "Fireside Induction," 5 U. ARK. LITTLE ROCK L. J. 1

(1982) ... 4, 10, 25

*iv Bronson, Does the Exclusion of Scrupled Jurors in Capital Cases Make the Jury More

Likely to Convict? Some Evidence from California, 3 WOODROW WILSON L.J. 11 (1980) ... 16,

17

Bronson, On the Conviction Proneness and Representativeness of the Death-Qualified Jury:

An Empirical Study of Colorado Veniremen, 42 U. COLO. L. REV. 1 (1970) ... 16, 17

J. CARLSMITH, P. ELLSWORTH & E. ARONSON, METHODS OF RESEARCH IN SOCIAL PSYCOLOGY (1976)

... 8

Cowan, Thompson & Ellsworth, The Effects of Death Qualification on Jurors' Predisposition

to Convict and on the Quality of Deliberation, 8 L. HUM. BEHAV. 53 (1984) ... 9, 10, 11,

12, 18, 20, 28

W. DANIELS, BIOSTATISTICS: A FOUNDATION FOR ANALYSIS IN THE HEALTH SCIENCES (1983) ...


1985 WL 669161 (U.S.)









23

A. DELAUNOIS (ed.), BIOSTATISTICS IN PHARMACOLOGY (1979) ... 23

B. EASTMAN, INTERPRETING MATHEMATICAL ECONOMICS AND ECONOMETRICS (1985) ... 23

Fitzgerald & Ellsworth, Due Process vs. Crime Control: Death Qualification and Jury

Attitudes, 8 L. & HUM. BEHAV. 31 (1984) ... 16, 17, 18, 19, 20, 28

Gillers, Deciding Who Dies, 129 U. PA. L. REV. 1 (1980) ... 5

Goldberg, Attitude Toward Capital Punishment And Behavior As A Juror In Simulated Cases

(Morehouse College, undated) ... 6, 7, 8

Goldberg, Toward Expansion of Witherspoon: Capital Scruples, Jury Bias, and Use of

Psychological Data to Raise Presumptions in the Law, 5 HARV. C.R.-C.L. L. REV. 53 (1970)

... 6, 7, 8

I. GUTTMAN, INTRODUCTORY ENGINEERING STATISTICS (3d ed. 1983) ... 23

Haney, Examining Death Qualification, 8 L. & HUM. BEHAV. 133 (1984) ... 15, 26

Haney, On the Selection of Capital Juries: The Biasing Effects of the Death-Qualification

Process, 8 L. & HUM. BEHAV. 121 (1984) ... 14, 26

*v Haney, Juries and the Death Penalty: Readdressing the Witherspoon Question, 26 CRIME

& DELINQ. 512 (1980) ... 15, 26

R. HASTIE, S. PENROD & N. PENNINGTON, INSIDE THE JURY (1983) ... 12, 13, 20

G. JACKSON, DECISIONS UNDER CERTAINTY: DRILLING DECISIONS BY GAS AND OIL OPERATORS (1979)

... 23

P. JEDAMUS, STATISTICAL ANALYSIS FOR BUSINESS DECISIONS (1976) ... 23

Jurow, New Data on the Effect of a "Death Qualified" Jury on the Guilt Determination

Process, 84 HARV. L. REV. 567 (1971) ... 9, 10, 16, 27

Kadane, After Hovey: A Note Taking into Account Automatic Death Penalty Jurors, 8 L. &

HUM. BEHAV. 115 (1984) ... 28

Kadane, Juries Hearing Death Penalty Cases: Statistical Analysis of a Legal Procedure, 78

J. AM. STAT. A. 544 (1983) ... 28

H. KALVEN & H. ZEISEL, THE AMERICAN JURY (1966) ... 13

Konecni & Ebbesen, Social Psychology and the Law: The Choice of Research Problems,

Settings, and Methodology in THE CRIMINAL JUSTICE SYSTEM: A SOCIAL-PSYCHOLOGICAL


1985 WL 669161 (U.S.)



ANALYSISSS 27 (V. Konecni & E. Ebbesen eds. 1982) ... 25

R. KUBO, STATISTICAL PHYSICS (1985) ... 23

Louis Harris & Associates, Inc., STUDY NO. 2016 (1971) ... 9, 11, 17, 19

Louis Harris & Assoc., Inc., STUDY NO. 814022 (1981) ... 27

R. MEAD & R. CURNOW, STATISTICAL METHODS IN AGRICULTURE AND EXPERIMENTAL BIOLOGY (1983)

... 23

Monahan & Walker, Social Authority: Obtaining, Evaluating, and Establishing Social

Science in Law, 133 U. PA. L. REV. (in press) ... 27

*vi Moran & Comfort, Neither "Tentative" nor "Fragmentary": Verdict Preference of

Impaneled Felony Jurors as a Function of Attitude Toward Capital Punishment, 71 J. APPLIED

PSYCHOL. (in press) ... 24, 25, 27

E. MURPHY BIOSTATISTICS IN MEDICINE (1982) ... 23

Oberer, Does Disqualification of Jurors for Scruples Against Capital Punishment

Constitute Denial of Fair Trial on Issue of Guilt? 39 TEXAS L. REV. 545 (1961) ... 6

Osser & Bernstein, The Death-Oriented Jury Shall Live, 1 U. SAN FERN. V. L. REV. 253

(1968) ... 15

PRECISION RESEARCH STUDY (1981) ... 17, 19

M. SAKS, JURY VERDICTS (1977) ... 13

Smith, A Trend Analysis of Attitudes Toward Capital Punishment, 1936-1974 in STUDIES OF

SOCIAL CHANGE SINCE 1948 (J. Davis ed. 1976) ... 20

Snyder, When Believing Means Doing: Creating Links Between Attitudes and Behavior in 2

VARIABILITY AND CONSISTENCY OF SOCIAL BEHAVIOR: THE ONTARIO SYMPOSIUM (M. Zanna, C.

Herman. E. Higgins eds. 1982) ... 24

Thompson, Cowan, Ellsworth & Harrington, Death Penalty Attitudes and Conviction

Proneness: The Translation of Attitudes into Verdicts, 8 L. & HUM. BEHAV. 95 (1984) ... 21

Vidmar & Ellsworth, Public Opinion and the Death Penalty, 26 STAN. L. REV. 1245 (1974)

... 20

White, The Constitutional Invalidity of Convictions Imposed by Death-Qualified Juries, 58

CORNELL L. REV. 1176 (1973) ... 11

Wilson, Belief In Capital Punishment And Jury Performance (Univ. of Chicago 1967) ... 6,

7, 8


1985 WL 669161 (U.S.)


http://www.westlaw.com/Find/Default.wl?rs=dfa1.0&vr=2.0&DocName=26STANLREV1245&FindType=Y

Zeisel, Some Insights Into The Operation of Criminal Juries (Univ. of Chicago 1957) ...

6, 7, 8, 24, 25, 27

H. ZEISEL, SOME DATA ON JUROR ATTITUDES TOWARD CAPITAL PUNISHMENT (1968) ... 6, 7, 8, 24,

25, 27

*1 BRIEF FOR AMICUS CURIAE AMERICAN PSYCHOLOGICAL ASSOCIATION IN SUPPORT OF

RESPONDENT

INTEREST OF AMICUS CURIAE

The American Psychological Association (APA), a non-profit, scientific, and professional

organization founded in 1892, is the major association of psychologists in the United

States. The APA has more than 60,000 members, including the vast majority of psychologists

holding doctoral degrees from accredited universities in the United States. Its purpose,

as reflected in its Bylaws, is to "advance psychology as a science and profession, and as

a means of promoting human welfare."

A central issue in this case is the applicability and methodological rigor of the social

science evidence introduced and considered in the courts below. Experimental social

psychologists have generated almost all the significant research pertinent to these issues

and provided most of the expert testimony introduced in this and similar cases. A

substantial number of APA's members are concerned with these issues, and more broadly,

with the usefulness of experimental psychology to the legal system.

The APA contributes amicus briefs only where it has special knowledge to share with the

Court; it has felt it particularly important to do so where scientific issues have been in

the forefront. E.g., Mills v. Rogers, 457 U.S. 291 (1982) (risks and benefits of

psychotropic medication); Metropolitan Edison Co. v. People Against Nuclear Energy, 460

U.S. 766 (1983) (measurability of psychological harm in potential disasters). The APA

regards this as one of those cases, and as an opportunity to offer this Court an objective

analysis of the social science evidence germane to a thoughtful resolution of the serious

constitutional questions confronting it.

*2 Petitioner and respondent have consented to the filing of this amicus brief. Their

letters of consent are on file with the Clerk of this Court.

INTRODUCTION AND SUMMARY OF ARGUMENT

Among the 270 exhibits introduced in the habeas corpus proceedings below were a wide

variety of studies generated primarily by social psychologists specializing in research

designed to address legal questions. The court heard from several social science experts

as well. These researchers and experts have addressed the empirical question whether a

"death-qualified" jury (one formed under standards developed in Witherspoon v. Illinois,

391 U.S. 510 (1968)), [FN1] is more likely to convict a capital criminal defendant than a

jury that-like juries formed in all non-capital criminal cases-also includes members of

the venire who because of absolute scruples against imposing the death penalty are

excluded from death-qualified juries.


1985 WL 669161 (U.S.)







FN1. With respect to the impact of Adams v. Texas, 448 U.S. 38 (1980), and

Wainwright v. Witt, 105 S. Ct. 844 (1985), on the social science research findings

relevant to this case, see infra, pp. 29-30.

In Witherspoon, this Court declined to rule that death-qualified juries are "less than

neutral with respect to guilt," id. at 520 n.18, and therefore unconstitutional, because

the research and data then extant was, in its view, too "tentative and fragmentary." Id.

at 517. The Court expressly left open the possibility it would rule differently if further

research more clearly demonstrated death-qualified juries' non-neutrality. Since that

time, researchers, particularly psychologists, have conducted new research designed to

answer definitively the question whether death-qualified juries are "conviction prone."

In ruling that excluding jurors who have absolute scruples against imposing the death

penalty at the guilt/innocence (culpability) phase [FN2] violates the Constitution, *3 the

courts below relied heavily on this research. In urging affirmance of the Eighth Circuit's

decision, respondent does as well. The States [FN3] attack the research on several

grounds, and at petitioner's urging the integrity of the data is one of the major

questions this Court has agreed to answer. See Petition for Writ of Certiorari at 20; 54

U.S.L.W. 3223 (U.S. October 8, 1985).

FN2. Not at issue in this case is the State's authority to exclude such jurors

during the sentencing phase. Also not at issue is the State's authority to exclude

jurors during both the culpability and sentencing phases whose "attitude toward the

death penalty would prevent them from making an impartial decision as to the

defendant's guilt." Witherspoon v. Illinois, 391 U.S. at 522-523 n.21.

FN3. By "States," APA refers to Petitioner Lockhart, as well as 16 States that

submitted a Brief Amici Curiae in support of petitioner.

To aid this Court in resolving this conflict, the APA will abstract and critique the

methodology and major empirical findings in the relevant research. Amicus concludes that

the extant research addresses the Sixth and Fourteenth Amendment issues that are at the

heart of this controversy. As amicus demonstrates, without credible exception, the

research studies show that death-qualified juries are prosecution prone, unrepresentative

of the community, and that death qualification impairs proper jury functioning. Point I.

Amicus next evaluates the data in light of the States' eight major criticisms of this

research. APA concludes that the States' objections are either mistaken or unrelated to

the relevant research. Further, amicus concludes that the research clearly satisfies the

criteria for evaluating the methodological soundness, reliability, and utility of

empirical research. Point II.

In sum, the research now addresses and answers the empirical question left open in

Witherspoon. Insofar as those data help resolve the constitutional questions at issue

here, they support the position of the respondent and the decision below.


1985 WL 669161 (U.S.)




http://www.westlaw.com/Find/Default.wl?rs=dfa1.0&vr=2.0&DB=780&FindType=Y&ReferencePositionType=S&SerialNum=1968131210&ReferencePosition=522


*4 ARGUMENT

I. THE SOCIAL SCIENCE DATA TENDERED BY RESPONDENT DEMONSTRATE THAT DEATH-QUALIFIED

JURIES ARE MORE PROPROSECUTION AND UNREPRESENTATIVE THAN TYPICAL CRIMINAL JURIES AND THAT

DEATH QUALIFICATION IMPAIRS JURY FUNCTIONING.

A. The Social Science Research has Focused on the Relevant Categories of Prospective

Jurors, Including that Subset of Jurors Excludable Under Witherspoon.

Understanding the relevant research first requires familiarity with the categories of the

venire generally used by researchers and judges addressing the issues before this Court.

The members of the venire competent to act as jurors can be aligned along a spectrum of

attitudes concerning the imposition of the death penalty. This continuum may be usefully

divided into the following classifications: [FN4]

FN4. This classification is adapted from Hovey v. Superior Court, 28 Cal. 3d 1, 616

P.2d 1301, 1311, 168 Cal. Rptr. 128 (1980). It was adopted in substantially similar

form by the district court below. Grigsby v. Mabry, 569 F. Supp. 1273, 1288 (E.D.

Ark. 1983), aff'd, 758 F.2d 226 (8th Cir. 1985). See Berry, Death Qualification and

the "Fireside Induction," 5 U. ARK. LITTLE ROCK L. J. 1, 2 (1982) [hereinafter

Berry].

Automatic Death Penalty Group (ADP)-These jurors will always vote for the death penalty

in a capital case.

Favor Death Penalty Group (FDP)-These jurors favor the death penalty but will not vote to

impose it in every capital case.

Indifferent Group-These jurors neither favor nor oppose the death penalty.

Oppose Death Penalty Group (ODP)-These jurors either oppose the death penalty or have

doubts about it but nevertheless will sometimes vote to impose it.

*5 Automatic Life Imprisonment Group (ALI)-These jurors will always vote for life

imprisonment rather than impose the death penalty. They may be challenged for cause at the

sentencing phase under the standards set forth in Witherspoon and are characteristically

known as "Witherspoon Excludables" (WEs).

WEs are further subdivided into two subsets, distinguished on the basis of an evaluation

of their impartiality at the culpability phase. One subset is not impartial; its members

state that they cannot be fair and impartial in deciding guilt, knowing that a guilty

verdict might ultimately result in imposition of the death penalty. These members of the

venire are characteristically described as "Nullifiers." See Wainwright v. Witt, 105 S.

Ct. at 852; Adams v. Texas, 448 U.S. at 48; Witherspoon v. Illinois, 391 U.S. at 522-523.

The second WE subset consists of those who, although they would never vote to impose the

death penalty, state they can be impartial on the issue of the defendant's guilt or

innocence. These members of the venire are characteristically denominated as "Guilt Phase


1985 WL 669161 (U.S.)












Includables" (GPIs). But because almost all American jurisdictions use the same jury to

decide both culpability and sentence in capital cases, see Gillers, Deciding Who Dies, 129

U. PA. L. REV. 1, 101-129 (1980), GPIs are excluded prior to the culpability phase along

with Nullifiers.

B. Social Science Research Conducted Over the Course of Three Decades Directly Addresses

the Constitutional Issues at Stake.

The empirical question whether death-qualified juries systematically favor the

prosecution at the culpability phase bears upon three distinct but related constitutional

doctrines under the Sixth and Fourteenth Amendments. [FN5] First, the process of death

qualification may produce juries that are less than neutral with respect to guilt and

uncommonly prone to convict. E.g., *6Witherspoon v. Illinois, 391 U.S. 510 (1968). This

relates to the outcome and verdict of the jury. Second, the systematic exclusion from the

jury of a group of eligible citizens possessing distinctive and relevant attitudes and

perspectives diminishes the representativeness of capital juries, and may implicate the

fair cross-section requirement. E.g., Duren v. Missouri, 439 U.S. 357 (1979); Taylor v.

Louisiana, 419 U.S. 522 (1975). This relates to the selection and demographic and

attitudinal composition of the jury. Third, the exclusion of these distinct perspectives

may impair proper functioning of the jury by reducing diversity of jurors, lessening the

accuracy of jury decisionmaking, and diminishing the potential for the counterbalancing of

community viewpoints. E.g., Ballew v. Georgia, 435 U.S. 223 (1978). This relates to jury

deliberation and adjudication.

FN5. APA will leave to the parties an extended discussion of the case law that

develops these doctrines.

1. The data show that death-qualified juries are conviction prone.

The question whether a jury qualified to carry out capital sentencing is more apt to

convict than juries in other criminal cases was first systematically analyzed and answered

in the affirmative in a seminal article by Oberer in 1961. [FN6] The article prompted

Witherspoon's use of three then-unpublished studies in his 1968 challenge to his death

sentence. These Witherspoon studies, subsequently discussed in this Court's decision in

his case, were authored by Wilson, Goldberg, and Zeisel [FN7] and are abstracted in Table

1.

FN6. Oberer, Does Disqualification of Jurors for Scruples Against Capital Punishment

Constitute Denial of Fair Trial on Issue of Guilt? 39 TEXAS L. REV . 545 (1961)

[hereinafter Oberer].

FN7. Wilson, Belief in Capital Punishment and Jury Performance (Univ. of Chicago

1967) (unpublished manuscript); Goldberg, Attitude Toward Capital Punishment and

Behavior as a Juror in Simulated Cases (Morehouse College, undated) (unpublished

manuscript); subsequently published as Goldberg, Toward Expansion of Witherspoon:

Capital Scruples, Jury Bias, and the Use of Psychological Data to Raise Presumptions


1985 WL 669161 (U.S.)







in the Law, 5 HARV. C.R.- C.L. L. REV V . 53 (1970); Zeisel, Some Insights into the

Operation of Criminal Juries (Univ. of Chicago 1957) (unpublished manuscript);

subsequently published as H. ZEISEL, SOME DATA ON JUROR ATTITUDES TOWARD CAPITAL

PUNISHMENT (Center for Studies of Criminal Justice. Univ. of Chicago Law School,

1968) [hereinafter Zeisel].

Table 1. Studies of Conviction-Proneness Before the Court in Witherspoon

Author Jurors Stimulus Death Results Sig. Level

Materials Qualifying [FNa]

Question

-------------------------------------------------------------------------------

Wilson 248 college Written des- ODP & WE Jurors p. < .02

[FNb] students criptions favoring

DP more

likely to

convict

Goldberg 200 college Written des- ODP & WE As above p. < .06

[FNc] students criptions

Zeisel 264 actual Actual Scruples As above p. < .04

[FNd] jurors felony

trials

against LP

-------------------------------------------------------------------------------

FNa. This category measures statistical significance. "p" value of .01 is

highly significant; .05 is significant; .10 is marginally significant. For

full discussion of statistical significance see infra, pp. 22-23.

FNb. Jurors were asked to agree/disagree with 15 statements measuring bias

toward prosecution and were given 5 descriptions of murder cases involving 6

defendants 17% scrupled jurors (conscientious scruples against death penalty)

voted for guilt in 5-6 cases compared to 30% of nonscrupled jurors. Jurors

favoring death penalty also more likely to assign severe punishments, more

likely to agree with pro-prosecution statements, and more likely to reject

insanity defense.

FNc. 100 whites and 100 blacks (116 male; 84 female), read synopses of 16

capital cases. Nonscrupled jurors voted to convict more often than scrupled

jurors (75% vs. 69%) and imposed more severe sentences. Blacks more likely to

oppose imposition of death penalty (76%) than whites (47%).

FNd. On their last day of jury service, jurors asked to disclose their jury's

first ballot vote and their own first ballot vote, yielding 464 split first

ballot votes. Weight of evidence controlled by dividing data into 11

different "constellations" of guilty/not guilty splits on first ballots. A


1985 WL 669161 (U.S.)


"constellation" was defined by how many jurors voted "guilty" on a particular

ballot. Nonscrupled jurors reached point of "equal likelihood" of voting

either guilty or not guilty when strength of evidence yielded 4 first ballot

guilty votes. Scrupled jurors did not exceed "equal likelihood" point until

there were 8 first ballot guilty votes. In 10 of 11 first ballot

constellations, nonscrupled jurors voted guilty more often than scrupled

jurors.

*8 lesser extent, they possessed at least one of four methodological problems that made

them, standing alone, less than definitive in guiding this Court's resolution of the

constitutional issues before it. First, the short written descriptions that Wilson and

Goldberg used as stimulus materials were not sufficiently realistic to permit confident

generalization to the real world. [FN8] Second, the participants in the Wilson and

Goldberg studies were college students. Although many of them may have been eligible to be

jurors, the extent to which students are representative of real jurors is not clear.

Third, the student jurors voted for guilt innocence without group deliberation. Fourth,

all three studies identified participants by the legally appropriate standards used at the

time. Because Witherspoon had not yet articulated the proper standard, the researchers

asked whether a jury formed by excluding all jurors having scruples against the death

penalty was more likely to convict than a jury comprised of those not having such

scruples. The difference in conviction behavior between those groups of jurors does not

directly address the Witherspoon issue because the scrupled group includes those opposed

to the death penalty (ODPs) and both subsets of WEs (Nullifiers and GPIs), rather than

consisting solely of GPIs.

FN8. See J. CARLSMITH, P. ELLSWORTH & E. ARONSON, METHODS OF RESEARCH IN SOCIAL

PSYCHOLOGY (1976).

Zeisel's study was open only to the last criticism. Furthermore, because it used actual

jurors in actual trials, Zeisel demonstrated that the differences in conviction rates

shown in other studies extend to actual situations, a demonstration that heightens the

reliability of results gleaned from simulations. And despite these criticisms, ??

*9 The Witherspoon Court's reluctance to rely on these studies was supported by more than

these methodological problems. Throughout the appeals process, Witherspoon's counsel had

contended that "the prosecution-prone character of 'death qualified' juries presented

'purely a legal question."' Witherspoon v. Illinois, 391 U.S. at 517 n.11. He proffered

these three studies for the first time in his brief to this Court. Thus, they had not been

"subjected to the traditional testing mechanisms of the adversary process." Ballew v.

Georgia, 435 U.S. 223, 246 (1978) (Powell, J., concurring).

In that light, the Witherspoon Court stated it could "only speculate . . . as to the

precise meaning of the terms used in those studies, the accuracy of the techniques

employed, and the validity of the generalizations made." 391 U.S. at 517 n.11. The Court


1985 WL 669161 (U.S.)







concluded that "[t]he data adduced by the petitioner . . . are too tentative and

fragmentary to establish that jurors not opposed to the death penalty tend to favor the

prosecution in the determination of guilt." Id. at 517. Although it declined to rely on

these data, the Court invited "a defendant convicted by [a death-qualified] jury . . . in

some future case [to] attempt to establish that the jury was less than neutral with

respect to guilt." Id. at 520 n.18. Although writing separately, Justice White shared this

view. Id. at 541 n.1 (White, J., dissenting).

Social scientists have responded to Witherspoon. They have produced research involving

participants classified by more precisely stated, legally-relevant Witherspoon standards,

under conditions that more closely approximated the real-life setting of the courtroom.

This research, which we label in Table 2 as the post-Witherspoon behavior studies,

included work by Jurow; Harris; and Cowan, Thompson & Ellsworth. [FN9]

FN9. Jurow, New Data on the Effect of a "Death Qualified" Jury on the Guilt

Determination Process, 84 HARV. L. REV. 567 (1971); Louis Harris & Associates, Inc.,

STUDY No. 2016 (1971) (unpublished manuscript), reported in White, The

Constitutional Invalidity of Convictions Imposed by Death-Qualified Juries, 58

CORNELL L. REV. 1176, 1178 n.12, 1185, 1194 (1973) and discussed in Hovey v.

Superior Court, 616 P.2d 1301, 1321 (1980); Cowan, Thompson & Ellsworth, The Effects

of Death Qualification on Jurors' Predisposition to Convict and on the Quality of

Deliberation, 8 L. HUM. BEHAV . 53 (1984) [hereinafter Cowan et al.].

Cowan et al. and several other studies to be discussed, infra, come from a special

issue on death qualification in Law and Human Behavior, the journal of the American

Psychology-Law Society, an APA division. For the convenience of the Court, nine

copies of the journal have been lodged with the Clerk of the Court.

Table 2. Post-Witherspoon Behavioral Studies of Conviction-Proneness

Author Jurors Stimulus Death Results Sig. Level

Materials Qualifying [FNa]

Question

-------------------------------------------------------------------------------

Jurow 211 persons, 2 audiotapes Appropriate Jurors "not p. < .01

[FNb]

1/3 former of simulated WE questions opposed" to (Case I)

DP

jurors murder trials more likely p. < .10

to convict (Case II)

Harris Nationwide Written desc- Appropriate As above p. < .0001

[FNc] riptions

random scriptions WE questions

sample

of 2068


1985 WL 669161 (U.S.)








adults

Cowan et 288 jury-el- Videotape of Precise WE, As above p. < .01

al. igible

[FNd]

murder trial GPI

questions

& former

jurors

-------------------------------------------------------------------------------

FNa. See Table 1, note a.

FNb. Audiotapes included opening statements, examination of witnesses,

closing arguments, jury instructions. Although Jurow found Case II only

marginally statistically significant. Berry, supra note 4 at 20 nn.44, 45,

argues that differences in both of Jurow's cases were statistically

significant. See also Cowan et al. at 58, who used the same cases, and show

that 44.7% of death-qualified jurors (DQs) vote to convict in Case I as

compared with 33.3% of WEs; in Case II, 60% of DQs vote to convict compared

to 42.9% of WEs.

FNc. In face-to-face interviews, jurors given 4 descriptions of criminal

cases. In each case, DQs voted to convict more often than ALIs. Overall, DQs

voted to convict in 63% of cases compared to 56% for WEs. DQs significantly

more willing to ignore procedural safeguards to vote for conviction than

ALIs.

FNd. See discussion in text.

*11 The Jurow and Harris studies are subject to legitimate reservations. The WEs in both

groups may have included Nullifiers. Because Nullifiers plainly may be excluded under the

rule in Witherspoon, the failure to distinguish them may have created an overinclusive

group. Moreover, neither Jurow nor Harris examined jury, as well as juror, behavior.

These two methodological gaps-whether GPIs differ in conviction rates from those who are

death-qualified under Witherspoon, and whether the findings concerning juror behavior

survive deliberation by juries-were filled in by Cowan et al., the most sophisticated

conviction-rate study to date. Participants were told to assume they were being called as

jurors in a first degree murder trial and that the judge would ask them about their

attitudes toward the death penalty. Each participant then was asked whether as a juror

he/she would be unwilling to consider voting to impose the death penalty in any case.

Those who answered unequivocally in the affirmative were classified as WEs, in the

negative as death-qualified jurors (DQs). Then the participants were asked whether in the

culpability phase they could follow the judge's instructions and decide the question of

guilt or innocence in a fair and impartial manner based on the evidence and the law. Those

who said they could not be fair and impartial in deciding culpability knowing that *12


1985 WL 669161 (U.S.)


conviction might lead to the death sentence were classified as Nullifiers and excluded

from the study. In this manner, 258 participants were identified as DQs and 30 as GPIs.

These remaining 288 participants, 104 of whom had actually served on juries, viewed a two

and one-half hour videotaped reenactment of an actual murder trial. [FN10] The judge and

the attorneys, portrayed by an actual judge and two experienced criminal lawyers, read the

entire original transcript of the trial and spontaneously improvised the reenactment. The

case was complex and afforded several plausible interpretations and verdict preferences

(first or second degree murder, voluntary manslaughter, accidental homicide,

self-defense). The tape was pretested and judged as "convincing and realistic." Cowan et

al., supra note 9, at 64.

FN10. The tape was prepared by Reid Hastie, Ph.D., an expert witness for respondent

in the trial below, who used the tape in his research on the jury deliberation

process. See R. HASTIE, S. PENROD & N. PENNINGTON, INSIDE THE JURY 47-51 (1983).

As soon as the taped presentation was completed, participants were asked to record how

they would vote. The vote was used as the basic measure of predisposition to convict. All

GPIs were then assigned to mixed juries composed of 8-10 DQs and 2-4 GPIs. These juries

were told to deliberate for one hour. The deliberations were recorded on audio- and

videotape. Jurors once again recorded their personal verdicts. This vote was used as the

measure of post-deliberation proneness to convict. The results are indicated in Table 3.

Table 3. Verdict Choices of DQ and GPI Jurors in Cowan et al.

Predeliberation Ballot Postdeliberation Ballot

-------------------------- --------------------------

Verdict DQ GPI DQ GPI

---------------------------------------------------------------------

1st Degree 7.8% 3.3% 1.0% 3.4%

2nd Degree 21.3% 23.3% 17.3% 13.8%

Manslaughter 48.9% 26.7% 68.0% 48.3%

Not Guilty 22.1% 46.7% 13.??% 34.5%

---------------------------------------------------------------------

*13 The difference between the DQ and GPI subjects on the pre-deliberation ballot are

significant beyond the .01 level (1 in 100), indicating that DQ subjects are more likely

than GPIs to vote guilty after viewing precisely the same trial. The post-deliberation

personal verdict preferences permitted the researchers to test the robustness of jurors'

initial verdict preferences after jury deliberation. The differences remained after

deliberation and continued to be significant at the .01 level. [FN11]

FN11. To assure that the results were not due to extraneous variables, the authors

used a statistical procedure to discover whether factors other than attitudes toward


1985 WL 669161 (U.S.)


the death penalty were associated with conviction rate. They found that age,

education, prior jury experience, and employment status were unrelated to their

behavior.

Two important findings emerge from the study, as inspection of Table 3 reveals. First,

GPIs take their duties seriously and will vote to convict. Over one-half of them voted to

convict of some degree of homicide prior to deliberation and an even greater proportion

(65%) voted to convict after deliberation in mixed juries. Second, after deliberation the

modal (most common) verdict among GPIs was manslaughter, the same as that for DQs,

although GPIs acquittal rate remained significantly higher.

A legitimate criticism that may be made of the Cowan et al. study is that none of the

juries deliberated to a verdict. Two factors attenuate any possible effects this factor

may have had on the ultimate conclusions. First, research shows that jurors' initial vote

preference predicts the jury's ultimate decision to a highly significant extent. [FN12]

Second, these data are not idiosyncratic. They comport with the findings from all other

studies, regardless of methodology, subjects, and stimulus materials.

FN12. H. KALVEN & H. ZEISEL, THE AMERICAN JURY 488 (1966); see R. HASTIE, S. PENROD

& N. PENNINGTON, INSIDE THE JURY (1983); M. SAKS, JURY VERDICTS (1977).

*14 The research discussed in this section shows that death qualified juries act more

favorably to the prosecution. In addition, there is a distinct and independent

proprosection effect created by death qualification during the voir dire process itself.

All capital jurors experience the process of death qualification. Because many prospective

jurors "may not know how they will react when faced with imposing the death sentence, or

may be unable to articulate, or may wish to hide their true feelings," Wainwright v. Witt,

105 S. Ct. 844, 852 (1985), the voir dire process may be quite lengthy. In many

jurisdictions voir dire occurs in the presence of other prospective jurors and can also be

highly repetitive. Haney [FN13] has studied the effects of voir dire on conviction

proneness.

FN13. On the Selection of Capital Juries: The Biasing Effects of the

Death-Qualification Process, 8 L. & HUM. BEHAV. 121 (1984) (lodged with the Court).

After WEs were excluded from the sample, Haney randomly assigned 67 jury-eligible adults

to one of two experimental conditions. They watched either a two-hour videotape of a

standard criminal voir dire including death qualification or an identical tape from which

the death-qualification portion had been deleted. At the conclusion of the tapes, all

subjects responded to a series of items designed to measure their attitudes and beliefs

about the case whose voir dire they had just observed.

Those exposed to the death-qualification voir dire were significantly more conviction

prone and were more likely to believe that the judge, the prosecution, and even the

defense attorneys thought the defendant was guilty. Haney also found disturbing evidence

of the effects of the death-qualifying voir dire on jurors' attitudes toward the


1985 WL 669161 (U.S.)




appropriate sentence. Of the 32 jurors who heard an ordinary voir dire, only seven said

that if the defendant were convicted of a capital crime, death was the appropriate

penalty. Of the 35 jurors exposed to the death-*15 qualifying voir dire, 20 said that

death would be the appropriate penalty.

In a parallel examination of actual capital voir dire, Haney [FN14] found that judges and

attorneys frequently lapsed into language even more prejudicial than that used in his

experiment. They used phrases that made the verdict seem a foregone conclusion, such as,

by the court: "When I instruct the jury at the end of this trial, I will outline in detail

the factors to be weighed in deciding whether to impose a death penalty," id. at 138; and

"There are two parts to this case," id. at 137; and by the prosecutor: "You know all [sic]

that you are going to have to go through with the second phase," id. at 138.

FN14. Examining Death Qualification, 8 L. & HUM. BEHAV. 133 (1984) (lodged with the

Court); see also Haney, Juries and the Death Penalty: Readdressing the Witherspoon

Question, 26 CRIME & DELINQ. 512 (1980).

In sum, a substantial, internally consistent body of research demonstrates that the

exclusion of jurors who cannot accept capital punishment but who can be fair and impartial

with respect to guilt, leaves a population of eligible jurors who are predisposed in favor

of the prosecution as compared to population of citizens who comprise typical criminal

juries. This basic conclusion is supported by three decades of accumulated research. It is

uncontradicted by the results of any credible empirical study. [FN15] Furthermore, a

recent, methodologically rigorous study demonstrates that the death-qualification process

during voir dire in capital cases creates further proprosecution biases.

FN15. The only study that disputes these findings is Osser & Bernstein, The

Death-Oriented Jury Shall Live, 1 U. SAN FERN. V. L. REV. 253 (1968). This study,

not in evidence in this case, was discussed in Hovey where the court indicated that

both defense and prosecution experts criticized its methodology. Hovey v. Superior

Court, 616 P.2d 1301, 1325 n.80 (1980).

*16 2. The data show barring of "Witherspoon Excludables" creates unrepresentative

juries, thereby implicating defendant's right to a jury composed from a fair cross-section

of the community.

The empirical evidence addresses three questions relevant to the Sixth Amendment

requirement that a jury be chosen from a fair cross-section of the community. First, the

evidence may yield information as to the size of each of the directly relevant groups,

most particularly the WEs and their subsets-Nullifiers and GPIs. Second, it may reveal

whether WEs share relevant attitudes, making them a distinct and cognizable subset of the

population. Third, the evidence can inform us whether the exclusion of WEs results in

underrepresentation of other distinct and cognizable groups on the jury, particularly

blacks and women.

a. Death qualification excludes a significanily large subset of the population.


1985 WL 669161 (U.S.)




Given the wide time-span over which the studies were done and the different

classifications used, no one number can define the proportion of WEs in the population,

but the data are uniformly within a rather narrow range of 10-17%, supporting their

reliability. [FN16] Two surveys, using *17 random samples and asking sophisticated

Witherspoon questions, found that 11% and 17%, respectively, of subjects are classifiable

as GPIs. [FN17] In Arkansas, where McCree was tried, a review of voir dire transcripts

revealed that WEs comprise 14% of venire members. [FN18] All of these data were offered by

respondent and subjected to cross-examination in the habeas proceedings below. [FN19]

FN16. Bronson, On the Conviction Proneness and Representativeness of the

Death-Qualified Jury: An Empirical Study of Colorado Veniremen, 42 U. COLO. L. REV .

1 (1970) [hereinafter Bronson I] (jury lists, 11% WEs, classification only

approximated Witherspoon standard); Bronson, Does the Exclusion of Scrupled Jurors

in Capital Cases Make the Jury More Likely to Convict? Some Evidence from

California, 3 WOODROW WILSON L. J. 11 (1980) [hereinafter Bronson II]) (author

replicated study in California with a more legally relevant Witherspoon question and

found 93% overlap of his "strongly opposed" group in Bronson I with WEs in Bronson

II); Jurow, supra note 9 (nonrandom sample, 10% WEs; data likely to underestimate

size of WEs in general population as author's sample 99% white and 80% male);

Fitzgerald & Ellsworth, Due Process vs. Crime Control: Death Qualification and Jury

Attitudes, 8 L. & HUM. BEHAV. 31 (1984) (random sample, 17% GPIs) [hereinafter

Fitzgerald & Ellsworth] (lodged with the Court); Arkansas Archival Study (1981)

(unpublished) cited at Grigsby v. Mabry, 758 F.2d 226, 234 (8th Cir. 1985) (voir

dire transcripts, 14% WEs); Precision Research Study (1981) (unpublished) discussed

by trial court at 569 F. Supp. at 1294 and appellate court at 758 F.2d at 233

(random sample, 11% GPIs).

FN17. Precision Research Study, supra note 16; Fitzgerald & Ellsworth, supra note

16.

FN18. Arkansas Archival Study, supra note 16.

FN19. After listening to the testimony, the district court indicated it would

"readopt[] what it stated in its first 'Grigsby' opinion concerning the size of the

excluded group, adding only that the evidence presented at the hearing after remand

reinforces the conclusion that the group [of WEs] is of substantial size both

nationally and within the state of Arkansas." Grigsby v. Mabry, 569 F. Supp. at

1285.

b. Those excluded through death qualification share common attitudes on issues related

to criminal justice.

Studies by Bronson (I & II), Harris, and Fitzgerald & Ellsworth [FN20] all demonstrate

that jurors' attitudes toward the death penalty are systematically related to their

ideological position on a broad spectrum of issues related to criminal justice. These


1985 WL 669161 (U.S.)






studies used different methods of defining subjects on the basis of their death penalty

attitudes, e.g., "scrupled" vs. "nonscrupled"; "strongly *18 oppose" vs. all others; and

some, the newest and most careful, compared DQs and WEs as defined by Witherspoon. All

yielded the same conclusion: Those who are more favorable to the death penalty share

attitudes toward criminal justice that are significantly more favorable to the

prosecution.

FN20. Bronson I, supra note 16 (718 persons drawn from jury lists asked 5

attitudinal items; participants favoring death penalty more likely to agree with

pro-prosecution items; sig. level = p. < .001); Bronson II, supra note 16 (755

members of venire in one survey and 707 in another interviewed regarding position on

death penalty showed that those favoring death penalty more likely to support

pro-prosecution positions); Harris, supra note 9; Fitzgerald & Ellsworth, supra note

16.

Fitzgerald & Ellsworth is the most methodologically rigorous of the studies and

exemplifies the nature of the attitudes investigated. The study was conducted under the

auspices of an independent research firm that interviewed 811 randomly selected,

jury-eligible participants. Interviewees were asked to indicate their position on the

death penalty, from "strongly favor" to "strongly oppose." DQs and WEs were classified in

precisely the same manner as in Cowan et al., supra note 9, thus removing Nullifiers and

permitting the researchers to compare DQs directly to GPIs. The 717 interviewees remaining

after Nullifiers were removed then responded to 13 items designed to tap, inter alia,

specific attitudes toward self-incrimination ("A person on trial who doesn't take the

stand and deny the crime is probably guilty"); inadmissible evidence ("If the police

obtain evidence illegally it should not be permitted in court, even if it would help

convict a guilty person"; "Despite the judge's instructions to ignore a confession

reported in the media but not in evidence, I would take the confession into consideration

since it clearly indicates the defendant's guilt"); as well as feelings about opposing

counsel ("District attorneys have to be watched carefully, since they will use any means

they can to get convictions"; "Defense attorneys have to be watched carefully, since they

will use any means to get their clients off"). [FN21]

FN21. These questions were carefully phrased. To reduce response bias and enhance

the reliability of results, the direction of the items were balanced, so that those

favoring prosecution and those favoring defense were in equal proportions.

On all 13 items, DQs answered in such a way as to indicate attitudes more favorable to

the prosecution than *19 GPIs. The differences in responses between the two groups were

significant at the .05 level or beyond with respect to 11 items (four were significant

beyond .001), marginally significant with respect to one item, and nonsignificant with

respect to one. Conversely, GPI's responses were significantly different from DQs'

responses on 11 of 13 questions, and less favorable to the prosecution on all 13.

Combining the results from all 13 items into a general index of prosecution-proneness, the

authors found a highly statistical significant difference in the proportion of


1985 WL 669161 (U.S.)


pro-prosecution attitudes between DQs and GPIs (p < .0001). See Hovey v. Superior Court,

616 P.2d 1301, 1337 (1980). Thus, the data show that GPIs share common attitudes on

relevant issues. [FN22]

FN22. The evidence concerning attitudinal similarity is germane not only to the fair

cross-representation issue but to the neutrality issue as well. The DQ process

creates not only a less diverse jury, but insofar as pro-prosecution attitudes are

related to conviction-proneness (and they appear to be), it creates a less than

neutral jury as well. The studies should be read in that light.

c. Death qualification results in underrepresentation on juries of blacks and women.

All the studies reporting relevant information [FN23] indicate that death qualification

has strong racial and gender effects, decreasing the proportion of blacks and women

eligible to serve as jurors during the culpability phase of the trial. Table 4 depicts the

percentage of WEs or GPIs by race and gender, as revealed in these studies. [FN24]

FN23. Harris, supra note 9; Fitzgerald & Ellsworth, supra note 16; Precision

Research Study, supra note 16.

FN24. See Hovey v. Superior Court, 616 P.2d 1301, 1337-1339 (1980) (collecting and

summarizing data). These data comport with the extensive findings of Zeisel, supra

note 7, which include data derived from three national Gallup polls conducted

between 1960-1965 and from other researchers as well. See Vidmar & Ellsworth, Public

Opinion and the Death Penalty, 26 STAN. L. REV. 1245 (1974); see also Smith, A Trend

Analysis of Attitudes Toward Capital Punishment, 1936-1974 in STUDIES OF SOCIAL

CHANGE SINCE 1948 (J. Davis ed. 1976). Other demographic variables such as age,

religion, and education are only weakly related to death penalty attitudes. See

Fitzgerald & Ellsworth, supra note 16, at 35.

Table 4. Percent of Cognizable Groups Classified as WEs or GPIs [FNa]

Author Black White Female Male

-------------------------------------------------------------------------------

Harris 46% 29% 37% 24%

Fitzgerald & Ellsworth 26% 16% 21% 13%

Precision Research 29% 9% 13% 8%

-------------------------------------------------------------------------------

FNa. In Harris, jurors classified as WEs; in other two, jurors classified as

GPIs.

*20 These indirect effects of death qualification on such cognizable groups pose

independent constitutional problems. See, e.g., Taylor v. Louisiana, 419 U.S. 522 (1975);

Peters v. Kiff, 407 U.S. 493 (1972).

3. The data suggest that death qualification interferes with the proper functioning of


1985 WL 669161 (U.S.)









the jury.

This Court has defined some of the attributes of a properly functioning jury-qualities

that make for effective jury deliberation. It has emphasized thorough, accurate, impartial

factfinding; robust debate; critical evaluation of testimony; "the counterbalancing of

various biases"; and the correct comprehension and application of rules of law and the

standard of reasonable doubt. [FN25] There is evidence that death qualification impairs

these attributes.

FN25. Ballew v. Georgia, 435 U.S. 223 (1978); Apodaca v. Oregon, 406 U.S. 404

(1972); Johnson v. Louisiana, 406 U.S. 356 (1972). See generally R. HASTIE, S.

PENROD, N. PENNINGTON, INSIDE THE JURY (1983).

Cowan et al., supra note 9, studied two types of juries, one comprised totally of DQs and

one comprised of 8-10 DQs and 2-4 GPIs (mixed). They found that those who had participated

on the mixed juries were able, to a significant extent, to remember more of the facts and

the evidence than those who served on DQ juries. Jurors who had participated in the mixed

juries were more critical of both defense and prosecution witnesses than those who had

served on the death-qualified juries.

*21 Thompson, Cowan, Ellsworth & Harrington [FN26] showed participants a videotape of the

conflicting testimony of a black defendant and a white police officer. To a highly

significant extent (p < .0002), the death-qualified jurors were more likely than the

excludable jurors to interpret the testimony as favorable to the prosecution. The

exclusion of GPIs, therefore, serious diminishes the opportunity for the presentation of

conflicting views.

FN26. Death Penalty Attitudes and Conviction Proneness: The Translation of Attitudes

into Verdicts, 8 L. & HUM. BEHAV. 95 (1984) (lodged with the Court).

In this same research report, the authors suggest another mechanism that affects the

proper functioning of the jury. They found that WEs felt that mistaken convictions were a

more regrettable error than mistaken acquittals, while DQs did not differentiate the two.

This indicates that WEs' attitudes may more properly fit the mandated asymmetry of the

reasonable doubt standard. [FN27]

FN27. Cf. Addington v. Texas, 441 U.S. 418 (1979); In re Winship, 397 U.S. 358

(1970).

Compared with the long history of research on neutrality and representativeness, the

research on the effects of death qualification is quite recent, and as yet there are few

studies. Nonetheless, the record is consistent, uncontroverted by the petitioner, and

indicates that adding the distinctive perspective of GPIs may improve the performance of

juries with respect to accurate factfinding, critical scrutiny of testimony, and the

proper application of the standard of reasonable doubt.

II. CONTRARY TO THE STATES' CRITICISMS, WHICH ARE EITHER ERRONEOUS OR UNRELATED TO THE


1985 WL 669161 (U.S.)








PERTINENT RESEARCH, THE SOCIAL SCIENCE DATA TENDERED BY RESPONDENT SATISFY APPLICABLE

CRITERIA FOR EVALUATING THE SOUNDNESS OF SCIENTIFIC RESEARCH.

The States' advance what amicus construes as eight criticisms of the studies relied upon

by respondent and *22 discussed above. [FN28] Amicus concludes that none of these

criticisms supplies a reason to disregard these studies.

FN28. APA believes these criticisms are a fair summary of the material contained in

the States' briefs. See Brief for Petitioner at 40-50; Brief for Amici Curiae at

23-30. Criticisms 1 & 2 are contained in both briefs; criticisms 3, 4, 5, 6, & 8

come primarily from the petitioner; criticism 7 is unique to the amici States.

1. Proof Based on Statistically Significant Findings is Used by All Sciences and Endorsed

by this Court. The issue the States stress most is that social science evidence

demonstrating that death-qualified juries are proprosecution depends upon statistical

significance as proof. Although the terminology is different, the underlying logic of

significance testing has been known to the law for centuries and is useful in resolving

empirical questions relevant to legal decisions, see Hazelwood Sch. Dist. v. United

States, 433 U.S. 299, 311 n.17 (1977); Castenada v. Partida, 430 U.S. 482, 496 n.17

(1977), as it is in all other fields that are concerned with realities, not just abstract

theories.

Significance levels, in fact, provide rigor and discipline for empirical researchers.

When faced with data (evidence) that is arguably ambiguous, statistical decision theory

requires that a "null hypothesis" be adopted as the starting point. The null hypothesis is

an assumption of "no differences," equivalent to the legal presumption that the

plaintiff's or prosecution's case is not proven until the evidence is submitted. The

presumption will prevail if the data are insufficient to permit its rejection.

A significance level is, thus, the statistical equivalent of a legal standard of proof.

It states the level of certainty the data must reach to reject the presumption of no

differences. In most fields, social science included, that standard is set by convention

at .05 or lower. In other words, if, after considering the evidence, the probability of an

erroneous rejection of the null hypothesis is less likely than 1 in 20, then the null

hypothesis can be rejected. This customary level of statistical significance *23 is

roughly equivalent to proof beyond a reasonable doubt. As such, it is far more reliable

than most evidence relied upon by the courts.

This Court has relied on tests of statistical significance to support the strength of

empirical evidence in cases of fundamental constitutional dimensions, particularly with

regard to juries. The Court has asked and answered empirical questions when deciding

whether a selection system has unconstitutionally excluded minorities from petit juries,

Alexander v. Louisiana, 404 U.S. 625 (1972); Carter v. Jury Commission, 396 U.S. 320

(1970), and grand juries, Castenada v. Partida, 430 U.S. 482 (1977). It has relied upon

empirical evidence in determining whether reduced size affects the process or product of


1985 WL 669161 (U.S.)








jury deliberation, Ballew v. Georgia, 435 U.S. 223 (1978); Colgrove v. Battin, 413 U.S.

149 (1973); Williams v. Florida, 399 U.S. 78 (1970), and whether nonunanimous juries

deliberate or decide differently from unanimous juries, Apodaca v. Oregon, 406 U.S. 404

(1972); Johnson v. Louisiana, 406 U.S. 356 (1972). In fact, tests of statistical

significance are fundamental to research and decisionmaking in every empirical field in

contemporary society. [FN29]

FN29. E.g.: Agriculture-R. MEAD & R. CURNOW, STATISTICAL METHODS IN AGRICULTURE AND

EXPERIMENTAL BIOLOGY (1983); Business-P. JEDAMUS, STATISTICAL ANALYSIS FOR BUSINESS

DECISIONS (1976); Economics-B. EASTMAN, INTERPRETING MATHEMATICAL ECONOMICS AND

ECONOMETRICS (1985); Engineering-I. GUTTMAN, INTRODUCTORY ENGINEERING STATISTICS (3d

ed. 1983); G. JACKSON, DECISIONS UNDER CERTAINTY: DRILLING DECISIONS BY GAS AND OIL

OPERATORS (1979); Medicine-W. DANIELS, BIOSTATISTICS: A FOUNDATION FOR ANALYSIS IN

THE HEALTH SCIENCES (1983); A. DELAUNOIS (ed.), BIOSTATISTICS IN PHARMACOLOGY

(1979); E. MURPHY, BIOSTATISTICS IN MEDICINE (1982); Physics-R. KUBO, STATISTICAL

PHYSICS (1985); and even law-D. BARNES, STATISTICS AS PROOF: FUNDAMENTALS OF

QUANTITATIVE EVIDENCE (1983).

2. The Studies Measure Actual Behavior of Jurors, Not Only Ttheir Attiudes. Contrary to

the States' assertion, the record is not limited to attitudinal studies. *24 Zeisel found

that death penalty attitudes, as revealed by a single question, predicted the votes of

jurors in actual felony trials. [FN30] All of the studies in Tables 1 and 2, supra,

measured actual votes in real or simulated cases, not just attitudes. Moreover, in all of

the studies, juror attitudes did predict juror behavior.

FN30. Zeisel's work has now been supported by Moran & Comfort, Neither "Tentative"

nor "Fragmentary": Verdict Preference of Impaneled Felony Jurors as a Function of

Attitude Toward Capital Punishment, 71 J. APPLIED PSYCHOL . (in press) [hereinafter

Moran & Comfort]. The authors, studying actual jurors serving on capital juries,

found a significant relationship between attitude toward death penalty and

conviction proneness. According to the study, those who strongly favor the death

penalty are more likely to be pro-prosecution, white, male, and politically

conservative.

There has been considerable controversy among social scientists about the general

correlation between attitudes and behavior. In some contexts, for example, where there is

strong cultural pressure to express certain attitudes, reported attitudes may be a poor

predictor. In others, such as election polls, they are an excellent predictor. The major

relevant findings are that attitudes are better predictors of behavior when people are

forced to behave one way or another and cannot simply leave the situation, and when the

relevance of the attitudes to the behavior is clear. [FN31] These conditions are met in

the jury context, which may explain the high correlation between attitude and behavior

observed in all studies.

FN31. See Abelson, Three Models of Attitude-Behavior Consistency in 2 VARIABILITY


1985 WL 669161 (U.S.)










AND CONSISTENCY OF SOCIAL BEHAVIOR: THE ONTARIO SYMPOSIUM 131 (M. Zanna, C. Herman,

E. Higgins eds. 1982) [hereinafter Ontario Symposium]; Snyder, When Believing Means

Doing: Creating Links Between Attitudes and Behavior in ONTARIO SYMPOSIUM.

3. The Findings are Not Based Solely on Simulations or Adversely Affected by

"Misclassification." Taken as a Whole, They are Demonstrably Reliable. The States claim

the data are unreliable because they are based on simulations. *25 This criticism is not

supportable. Two studies (Zeisel; Moran & Comfort) were based on post-deliberation

interviews with actual jurors. Moreover, the simulation studies in the record are not

unrealistic. A variety of stimulus materials have been used, ranging from written

descriptions of cases, to half-hour audiotapes, to a highly realistic two and one-half

hour videotape followed by a judge's jury instructions and jury deliberation. Moreover, as

the simulations in the studies become more realistic, the differences between DQs and WEs

become more pronounced. Apparently, any lack of realism in a given study has diluted the

magnitude of the observed differences. Without credible exception, the substantial body of

data supports the same conclusion. [FN32]

FN32. It would be theoretically possible to do a naturalistic experiment by

simultaneously trying a capital defendant before a variety of juries, e.g., some

containing DQs only, some with GPIs only, some with mixed DQs and GPIs, and then

assessing the relative tendency of these groups to convict or acquit. See Konecni &

Ebbesen, Social Psychology and the Law: The Choice of Research Problems, Settings,

and Methodology in THE CRIMINAL JUSTICE SYSTEM: A SOCIAL-PSYCHOLOGICAL ANALYSIS 27

(V. Konecni & E. Ebbesen eds. 1982). However, given three decades of consistent

findings with a wide variety of stimulus materials, such an experiment would be only

marginally useful. Moreover, this "ideal experiment" would raise serious practical,

legal and ethical problems. Berry, supra note 4, at 5-6.

The States' suggestions that researchers have misclassified some jurors as WEs or DQs and

that such misclassifications have created spurious differences between groupings of jurors

are likewise unsupported and untenable. The States complain particularly about studies

classifying jurors on the basis of a single death-qualification question rather than a

more extended voir dire. First, the method of classifying jurors has differed across

studies (including those conducted by Dr. Shure, an expert witness for the State, see

Brief for Petitioner at 44), with researchers varying the number and form of the questions

without producing substantial changes in the numbers of jurors classified as WEs, GPIs, or

DQs.

*26 Second, if "misclassification" did occur in any study, it would yield underestimates,

not overestimates, as implied by the States, of the true differences in conviction

proneness among different groupings of jurors. To the extent DQs are misclassified as WEs,

or vice versa, they will dilute the homogeneous character of each group, make the two

groups more like each other, and thereby obscure real differences.

Third, the research by Haney, supra notes 13 & 14, demonstrates that the lengthy death


1985 WL 669161 (U.S.)


qualification portion of voir dire adds a further prejudicial factor influencing death

qualified juries against the defendant. Thus, if any effects on verdicts were produced by

use of a full voir dire to classify jurors, it would be expected to increase, not

decrease, observed empirical differences.

4. Use of Diverse Methodologies Enhances the Reliability of the Research at Issue. The

States argue that the studies are suspect because researchers' use of different methods

reveals a "lack of consensus" vitiating the reliability of the evidence. In fact, the

variety of approaches used by investigators-all of which led to the same result-reinforces

their reliability.

The use of diverse subjects, stimulus materials, and empirical methods does not reveal a

"lack of consensus" but comports fully with the goal of "generalization," the accepted

rubric for evaluating how far beyond the specific facts of any one particular study one

can apply its findings. In making this evaluation, one must first consider whether the

findings can be generalized across persons, i.e., whether the subjects who participated in

the research differ in important ways from the people to whom the research is being

generalized. Next, one must consider whether the findings can be generalized across

settings, i.e., whether they apply in situations not directly involved in the study.

Finally, one must consider whether findings can be generalized over time. Thus, the

trustworthiness and generalizability of a study increases as independent investigators

arrive at a common conclusion. The more *27 often a study confirms prior research or is

confirmed by subsequent research and the more often a body of research with differing

methodologies supports a common proposition, the less likely it is that chance

fluctuations in the data or methodological anomalies account for the findings. [FN33]

FN33. See generally Monahan & Walker, Social Authority: Obtaining, Evaluating, and

Establishing Social Science in Law, 133 U. PA. L. REV. (in press).

The studies reported here meet these tests. Regardless of the decade in which the

research was done, the population studied, the stimulus materials used, the research

design employed (retrospective interviews with actual jurors, national public opinion

surveys, or increasingly sophisticated and controlled laboratory simulations), and

regardless of the contemporary state of public opinion about the death penalty, the

results have been the same-death-qualified jurors are more likely to convict than their

"excludable" counterparts.

5. The Findings Have Remained Remarkably Stable Over Almost 30 Years. The States charge

that the findings will change over time. The short answer is that the basic relationship

between attitudes toward the death penalty and conviction proneness have remained markedly

stable over three decades, from 1957 (Zeisel) to the present (Moran & Comfort). It is

especially notable that the relationship has been robust regardless of whether, at the

time the study was performed, the death penalty was popular or unpopular.

6. ADPs Are So Small a Group That Their Absence Has no Impact on the Validity of the


1985 WL 669161 (U.S.)


Findings. Studies reveal that the prevalence of jurors who would always vote for the death

penalty regardless of the evidence is exceedingly small, [FN34] ranging from 0.5% to 2%.

[FN35] Statistical *28 analysis based upon the data from these studies demonstrates that

"taking the automatic death penalty jurors into account will have little impact on the

findings" of Fitzgerald & Ellsworth and Cowan et al. [FN36]

FN34. See Louis Harris & Assoc., Inc., STUDY No. 814022 (1981) (0.7%); Arkansas

Archival Study, supra note 16 (0.5%); Jurow, supra note 9 (2%).

FN35. Gerald Shure, Ph.D., a psychologist expert who testified for the petitioner,

stated during the trial that in a telephone survey conducted in a wealthy area of

West Los Angeles he found 33.3% of his sample to be ADPs. But the trial court

criticized his methodology, indicated that Dr. Shure had himself acknowledged

potential errors and omissions in his study, and concluded that the percentage of

ADPs, both nationwide and in Arkansas, is negligible. 569 F. Supp. at 1307-1308. The

Eighth Circuit agreed. 758 F.2d at 237-238. This Court has also concluded that such

jurors "will be few indeed." Adams v. Texas, 448 U.S. 38, 49 (1980).

FN36. Kadane, After Hovey: A Note Taking into Account Automatic Death Penalty

Jurors, 8 L. & HUM. BEHAV. 115 (1984) (lodged with the Court); see Kadane, Juries

Hearing Death Penalty Cases: Statistical Analysis of a Legal Procedure, 78 J. AM.

STAT. A. 544 (1983).

A related criticism concerns the fact that the studies compared death-qualified and

excludable jurors, rather than death-qualified and typical criminal juries. Inevitably,

juries composed of a mixture of death-qualified and GPIs jurors will fall somewhere

between a jury composed solely of GPIs or DQs. Death-qualified jurors are more conviction

prone than such mixed juries would be. Thus, the size of the difference between the pure

WE and pure DQ groups in the research studies cannot be taken as an accurate estimate of

the size of the difference between jury verdicts, although the existence and direction of

the effect is clear from the studies. The precise magnitude of difference in a given case

will change as a result of many variables, e.g., quality of lawyering, first ballot

verdict, number of WEs on mixed juries, strength of the evidence.

The research demonstrates that the composition of juries in terms of death penalty

attitudes is an important variable and that over the long run eliminating GPIs will

increase the number of guilty verdicts. The closer the case, the more the attitudinal and

behavioral differences found between DQs and GPIs are crucial to the outcome. It is in

close cases where neutral juries comprised *29 of jurors representing a fair cross-section

of the community, who can critically and accurately evaluate the testimony and correctly

apply the law, are most essential. [FN37]

FN37. The States' criticism concerning proper comparisons does not in any way affect

the analysis of the data in regard to the Sixth Amendment requirements that juries

must function properly and be comprised of a fair cross-section of the community.


1985 WL 669161 (U.S.)






See pp. 16-21, supra.

7. It is Implausible that the Results are Products of Researcher Bias. There is no

support whatsoever for the States' claim that "researcher bias" produced the findings at

issue. Because the findings converge across a wide variety of methods and have emerged

from any laboratories, it is improbable that all the researchers share a common bias. In

at least two cases (Harris; Fitzgerald & Ellsworth) the data were gathered by

disinterested research firms whose livelihood depends on their reputations for

impartiality. Perhaps most importantly, the data presented here have withstood trial in

the twin crucibles of journal review by anonymous expert peers and the adversarial process

of cross-examination in a courtroom. These independent procedures, from science and law,

are designed to identify precisely the types of bias alleged by the States. [FN38]

FN38. See Barefoot v. Estelle, 463 U.S. 880, 901 (1983) ("We are unconvinced . . .

that the adversary process cannot be trusted to sort out the reliable from the

unreliable evidence").

8. Use of Adams/Witt, Rather than Witherspoon Standards Decreases the Death-Qualified

Jury's Neutrality, Representativeness, and Effectiveness. This Court's decisions in

Wainwright v. Witt, 105 S. Ct. 844 (1985), and Adams v. Texas, 448 U.S. 38 (1980),

clarified and broadened the permissible standard for exclusion of jurors developed in

Witherspoon. This modification, however, does nothing to undermine the integrity of the

findings presented here. In broadening the criteria for exclusion, Adams/Witt has the

effect of removing from the jury *30 some additional number of jurors who are disinclined

toward the death penalty, and transferring them from the DQ to the WE category. The

result, then, is to make the death-qualified juries under the Adams Witt rule even more

different from normal criminal juries than they were under Witherspoon, e.g., less

diverse, less able to properly evaluate the evidence, more pro-prosecution-all

exacerbating the effects of death qualification on the fact-finding process at the

culpability phase.

CONCLUSION

The data demonstrating that death-qualified juries are less than neutral with respect to

guilt, unrepresentative, and ineffective as compared to normal criminal juries are now

neither tentative nor fragmentary. The terms used in the relevant studies have been

precisely defined. The techniques employed have been carefully articulated. The stability

and convergence of the findings over three decades lend impressive support to their

validity. The studies of the past decade, particularly, have closely approximated the

real-life setting of the courtroom. Insofar as the social science data are relevant to the

resolution of the constitutional issues at stake in this case, therefore, amicus believes

they support affirmance of the decision below.

Respectfully submitted,

DONALD N. BERSOFF


1985 WL 669161 (U.S.)





(Counsel of Record)

DAVID W. OGDEN

ENNIS, FRIEDMAN, BERSOFF & EWING

1200 17th Street, N.W.,

Suite 400

Washington, D.C. 20036

(202) 775-8100

Attorneys for Amicus Curiae American Psychological Association

December 23, 1985

Lockhart v. McCree

1985 WL 669161

Briefs and Other Related Documents (Back to top)

• 1986 WL 727598 (Appellate Brief) Reply Brief of Petitioner (Jan. 06, 1986)

• 1985 WL 669163 (Appellate Brief) Brief of Amici Curiae (Dec. 24, 1985)

• 1985 WL 669162 (Appellate Brief) Brief of Amici Curiae (Dec. 23, 1985)

• 1985 WL 669165 (Appellate Brief) Brief for the National Center on Institutions and

Alternatives as Amicus Curiae (Dec. 23, 1985)

• 1985 WL 669160 (Appellate Brief) Brief of Petitioner (Nov. 25, 1985)

• 1985 WL 669159 (Appellate Brief) Brief of Amici Curiae, Alabama, Connecticut, Florida,

Georgia, Idaho, Illinois, Indiana, Mississippi, Missouri, New Hampshire, New Jersey, New

Mexico, Oregon, South Dakota, Texas, and Washington%n*%n (Nov. 22, 1985)

• 1985 WL 669158 (Appellate Brief) Reply Brief for Petitioner (Sep. 20, 1985)

END OF DOCUMENT


1985 WL 669161 (U.S.)







